This paper concerns the Generalized World Entities (GWEs) paradigm, an extension of the ordinary IoT/WoT (Internet of Things/Web of Things) approach. GWEs offer a unified way to seamlessly model i) conceptual representations of physical objects, humans, robots and low-level events and ii) higher level of abstractions corresponding to structured situations/behaviors implying mutual relationships among low level entities. The GWEs approach is currently implemented in terms of NKRL, the "Narrative Knowledge Representation Language", which is both a Knowledge Representation language and a Computer Science environment. It is expected to represent a significant contribution with respect to bridge the gap between the recognition of entities at sensor level and their description/management at conceptual level.